Please wait a minute...
Shopping Cart Email List Chinese
Advanced Search Fig/Tab
Frontiers of Computer Science

ISSN 2095-2228

ISSN 2095-2236(Online)

CN 10-1014/TP

Postal Subscription Code 80-970

2018 Impact Factor: 1.129

Featured Articles  |  Most Downloaded  |  Most Reading
Online Submission Subscribe to Our RSS Content Feed
Front. Comput. Sci.    2022, Vol. 16 Issue (5) : 165346    https://doi.org/10.1007/s11704-022-1213-7
LETTER
A Bayesian matrix factorization model for dynamic user embedding in recommender system
Kaihan ZHANG, Zhiqiang WANG(), Jiye LIANG, Xingwang ZHAO
Key Laboratory of Computational Intelligence and Chinese Information Processing of Ministry of Education, School of Computer and Information Technology, Shanxi University, Taiyuan 030006, China
 Download: PDF(1318 KB)   HTML
 Export: BibTeX | EndNote | Reference Manager | ProCite | RefWorks
Corresponding Author(s): Zhiqiang WANG   
Just Accepted Date: 16 February 2022   Issue Date: 22 April 2022
 Cite this article:   
Kaihan ZHANG,Zhiqiang WANG,Jiye LIANG, et al. A Bayesian matrix factorization model for dynamic user embedding in recommender system[J]. Front. Comput. Sci., 2022, 16(5): 165346.
 URL:  
https://academic.hep.com.cn/fcs/EN/10.1007/s11704-022-1213-7
https://academic.hep.com.cn/fcs/EN/Y2022/V16/I5/165346
Features 100k 10m Dianping
#users 943 69,878 37,081
#items 1,682 10,677 28,520
#ratings 100,000 10,000,054 1,077,845
density 6.3047% 1.3403% 0.1019%
Tab.1  Statistics of datasets
100k 10m Dianping
MAE RMSE MAE RMSE MAE RMSE
PMF 0.7414 0.9500 0.6125 0.7991 0.6531 0.8777
timeSVD++ 0.7318 0.9516 0.6163 0.8147 0.5999 0.7901
TMF 0.7676 0.9852 0.6118 0.8000 0.5599 0.7608
PCCF 0.7361 0.9351 0.6116 0.7970 0.5892 0.7740
BMFDE 0.7182 0.9177 0.6055 0.7904 0.5427 0.7225
Tab.2  The results compared with matrix factorization-based models
100k 10m Dianping
MAE RMSE MAE RMSE MAE RMSE
LightGCN 0.7234 0.9270 0.5993 0.7890 0.5312 0.7039
BMFDE 0.7182 0.9177 0.6055 0.7904 0.5427 0.7225
Tab.3  The results compared with neural network-based model
Fig.1  Training time (seconds) of the comparison methods on three datasets. (a) On 100k dataset; (b) on 10m dataset; (c) on dianping dataset
Fig.2  The influence of parameters s and d on MAE on three datasets. (a) On 100k dataset; (b) on 10m dataset; (c) on dianping dataset
1 Z Y Zhao , H Zhou , C Li , J Tang , Q T Zeng . Deepemlan: deep embedding learning for attributed networks. Information Sciences, 2021, 543: 382– 397
2 Y Koren. Collaborative filtering with temporal dynamics. In: Proceedings of the 15th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. 2009, 447– 456
3 C Y Zhang, K Wang, H K Yu, J L Sun, E P Lim. Latent factor transition for dynamic collaborative filtering. In: Proceedings of 2014 SIAM International Conference on Data Mining. 2014, 452– 460
4 G B Guo, F D Zhu, S L Qu, X W Wang. PCCF: periodic and continual temporal co-factorization for recommender systems. Information Sciences, 2018, 436: 56– 73
5 X N He, K Deng, X Wang, Y Li, Y D Zhang, M Wang. LightGCN: simplifying and powering graph convolution network for recommendation. In: Proceedings of the 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval. 2020, 639– 648
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed